[Qemu-devel] [PATCH] hw/arm/virt: add DT property /secure-chosen/sdtout-path being secure UART

[Qemu-devel] [PATCH] hw/arm/virt: add DT property /secure-chosen/sdtout-path being secure UART

[Jerome Forissier]

Bindings for /secure-chosen and /secure-chosen/stdout-path have been
proposed 1.5 years ago [1] and implemented in OP-TEE at the same time [2].

This patch creates the property when the machine is secure.

[1] https://patchwork.kernel.org/patch/9602401/
[2] https://github.com/OP-TEE/optee_os/commit/4dc31c52544a

Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
---
 hw/arm/virt.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/hw/arm/virt.c b/hw/arm/virt.c
index 0b57f87abc..a0faa40d64 100644
--- a/hw/arm/virt.c
+++ b/hw/arm/virt.c
@@ -712,6 +712,10 @@ static void create_uart(const VirtMachineState *vms, qemu_irq *pic, int uart,
         /* Mark as not usable by the normal world */
         qemu_fdt_setprop_string(vms->fdt, nodename, "status", "disabled");
         qemu_fdt_setprop_string(vms->fdt, nodename, "secure-status", "okay");
+
+        qemu_fdt_add_subnode(vms->fdt, "/secure-chosen");
+        qemu_fdt_setprop_string(vms->fdt, "/secure-chosen", "stdout-path",
+                                nodename);
     }
 
     g_free(nodename);
-- 
2.15.1

Re: [Qemu-devel] [PATCH] hw/arm/virt: add DT property /secure-chosen/sdtout-path being secure UART

[Peter Maydell]

On 5 October 2018 at 09:07, Jerome Forissier
<jerome.forissier@linaro.org> wrote:
> Bindings for /secure-chosen and /secure-chosen/stdout-path have been
> proposed 1.5 years ago [1] and implemented in OP-TEE at the same time [2].
>
> This patch creates the property when the machine is secure.
>
> [1] https://patchwork.kernel.org/patch/9602401/
> [2] https://github.com/OP-TEE/optee_os/commit/4dc31c52544a
>
> Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>

The patch seems OK, but it looks like the DT bindings are
not upstream yet? [*] I wouldn't want QEMU to start creating
bindings until the spec is definitely final.

[*] they're not in
https://www.kernel.org/doc/Documentation/devicetree/bindings/arm/secure.txt
anyway; is that the wrong place to look?

thanks
-- PMM

Re: [Qemu-devel] [PATCH] hw/arm/virt: add DT property /secure-chosen/sdtout-path being secure UART

[Jerome Forissier]

+CC: Rob Herring

On 10/05/2018 11:16 AM, Peter Maydell wrote:
> On 5 October 2018 at 09:07, Jerome Forissier
> <jerome.forissier@linaro.org> wrote:
>> Bindings for /secure-chosen and /secure-chosen/stdout-path have been
>> proposed 1.5 years ago [1] and implemented in OP-TEE at the same time [2].
>>
>> This patch creates the property when the machine is secure.
>>
>> [1] https://patchwork.kernel.org/patch/9602401/
>> [2] https://github.com/OP-TEE/optee_os/commit/4dc31c52544a
>>
>> Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
> 
> The patch seems OK, but it looks like the DT bindings are
> not upstream yet? [*]

No, they're not.

 I wouldn't want QEMU to start creating
> bindings until the spec is definitely final.

I understand, but I think we may have a chicken-and-egg problem, because
Rob said he'd like to see more usage of DT in secure world before
merging the bindings ;-)

That being said, I can certainly re-submit the binding to the LKML. Rob,
what do you say?

> 
> [*] they're not in
> https://www.kernel.org/doc/Documentation/devicetree/bindings/arm/secure.txt
> anyway; is that the wrong place to look?
> 
> thanks
> -- PMM
> 

Thanks,
-- 
Jerome

Re: [Qemu-devel] [PATCH] hw/arm/virt: add DT property /secure-chosen/sdtout-path being secure UART

[Jerome Forissier]

Hi Peter,

On 10/05/2018 11:26 AM, Jerome Forissier wrote:
> +CC: Rob Herring
> 
> On 10/05/2018 11:16 AM, Peter Maydell wrote:
>> On 5 October 2018 at 09:07, Jerome Forissier
>> <jerome.forissier@linaro.org> wrote:
>>> Bindings for /secure-chosen and /secure-chosen/stdout-path have been
>>> proposed 1.5 years ago [1] and implemented in OP-TEE at the same time [2].
>>>
>>> This patch creates the property when the machine is secure.
>>>
>>> [1] https://patchwork.kernel.org/patch/9602401/
>>> [2] https://github.com/OP-TEE/optee_os/commit/4dc31c52544a
>>>
>>> Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
>>
>> The patch seems OK, but it looks like the DT bindings are
>> not upstream yet? [*]
> 
> No, they're not.

Update: Rob has now accepted the bindings [3] (thanks!).

[3] http://patchwork.ozlabs.org/patch/980467/#2009642

-- 
Jerome

> 
>  I wouldn't want QEMU to start creating
>> bindings until the spec is definitely final.
> 
> I understand, but I think we may have a chicken-and-egg problem, because
> Rob said he'd like to see more usage of DT in secure world before
> merging the bindings ;-)
> 
> That being said, I can certainly re-submit the binding to the LKML. Rob,
> what do you say?
> 
>>
>> [*] they're not in
>> https://www.kernel.org/doc/Documentation/devicetree/bindings/arm/secure.txt
>> anyway; is that the wrong place to look?
>>
>> thanks
>> -- PMM
>>
> 
> Thanks,
> 

Re: [Qemu-devel] [PATCH] hw/arm/virt: add DT property /secure-chosen/sdtout-path being secure UART

[Peter Maydell]

On 15 October 2018 at 10:08, Jerome Forissier
<jerome.forissier@linaro.org> wrote:
> Hi Peter,
>
> On 10/05/2018 11:26 AM, Jerome Forissier wrote:
>> +CC: Rob Herring
>>
>> On 10/05/2018 11:16 AM, Peter Maydell wrote:
>>> On 5 October 2018 at 09:07, Jerome Forissier
>>> <jerome.forissier@linaro.org> wrote:
>>>> Bindings for /secure-chosen and /secure-chosen/stdout-path have been
>>>> proposed 1.5 years ago [1] and implemented in OP-TEE at the same time [2].
>>>>
>>>> This patch creates the property when the machine is secure.
>>>>
>>>> [1] https://patchwork.kernel.org/patch/9602401/
>>>> [2] https://github.com/OP-TEE/optee_os/commit/4dc31c52544a
>>>>
>>>> Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
>>>
>>> The patch seems OK, but it looks like the DT bindings are
>>> not upstream yet? [*]
>>
>> No, they're not.
>
> Update: Rob has now accepted the bindings [3] (thanks!).

OK; I've applied the QEMU patch to my target-arm.next branch.

thanks
-- PMM